THE PREBISCH-SINGER HYPOTHESIS: FOUR CENTURIES OF EVIDENCE
David I. Harvey, Neil M. Kellard, Jakob B. Madsen, Mark E. Wohar

This archive contains the main dataset employed along with four Gauss programs for computing the main results.

The raw data are in the file data2.txt. Each series is a relative commodity price index (unlogged), which are read into the Gauss programs (logarithms are taken within the Gauss programs). The data are listed annually from 1650 (some series begin with zeros as the data is not available until later), and the columns are in the following order:

ALUMINUM
BANANA
BEEF
COCOA
COFFEE
COPPER
COTTON
HIDE
JUTE
LAMB
LEAD
RICE
SILVER
SUGAR
TEA
TIN
TOBACCO
WHEAT
WOOL
ZINC
PIG IRON
COAL
NICKEL
GOLD
OIL

The four Gauss programs are as follows:

table1.gp: Produces the results of Table 1 of the paper.

table2.gp: Produces the 10%-level results of Table 2 of the paper. Results for the 5%- and 1%-levels are obtained by replacing the b_alpha value and the critical value in the code as appropriate.

table3.gp: Prodcues the 10%-level results of Table 3 of the paper. Results for the 5%- and 1%-levels are obtained by replacing the b_alpha value and the critical value in the code as appropriate.

table6.gp: Prodcues the results of Table 6 of the paper.

Results for Tables 4 and 5 are obtained by applying the table1.gp code to truncated series and the Grilli-Yang dataset.

